25

DISTILLATION
Keywords: Phase Equilibrium, Isothermal Flash, Adiabatic Flash, Batch Distillation

Distillation refers to the physical separation of a mixture into two or more fractions
that have different boiling points. The general objective is to separate substances having
different vapor pressures at any given temperature. If a liquid mixture of two volatile
materials is heated, the vapor that comes out with a higher concentration of the lower boiling
material than the liquid from which it was evolved. Conversely, if a warm vapor is cooled,
the higher boiling material has a tendency to condense in a greater proportion than the lower
boiling material. The early distillers of alcohol for beverages applied these fundamental ideas.

A distillation column consists of a series of plates (or trays). In normal operation, there is a
certain amount of liquid on each plate, and some arrangement is made for ascending vapors
to pass through the liquid and make contact with it. The descending liquid flows down from
the plate above through a downcomer, across the next plate, and then over a weir and into
another downcomer to the next lower plate as shown in fig 1. Earlier on, bubble caps were
used on the trays for the purpose of vapor and liquid contacting. Recent developments
include the use of sieve tarys, valve trays, perforated or ballast trays.

Fig. 25.1: Section of Distillation Column

Distillation columns can be as high as 200ft. Diameters as large as 44ft have been used;
operates at a pressure ranging from 15mm to 500psia. As indicated in fig 25.2, the overhead
vapor V1 , upon leaving the top plate enters the condenser where it is either partially or totally
condensed. The liquid formed is collected in an accumulator from which the liquid stream L 0
(reflux) and the top product stream D (distillate) are withdrawn. When the overhead vapor,
V1 is totally condensed to the liquid state to produce L0 and D is withdrawn as a liquid, the
condenser is called total condenser. If V1 is partially condensed to the liquid state to
produce L0 and D is withdrawn as a vapor, the condenser is called partial condenser. The
amount of reflux is generally expressed in terms of reflux ratio, L0/D. The liquid that leaves
the bottom plate of the column enters the reboiler, where it is partially vaporized. The vapor
produced is allowed to flow back up through the column, and the liquid is withdrawn from
the reboiler known as bottom product, B.

Fig. 25.2: Continuous Distillation Column

Fundamental Principles involved in distillation

In order to compute top product, D and bottom product, B ; it is necessary to obtain a solution
of equilibrium relationships, component material balances, total material balances and energy
balances respectively.

Consider first equilibrium relationships.

Physical equilibrium

A two phase multicomponent mixture is said to be in equilibrium if

The temperature Tv of the vapor phase is equal to Tl of the liquid phase.

The total pressure Pv throughout the vapor phase is equal to the total pressure P l
throughout the liquid phase.
The tendency of each component to escape from the liquid phase to the vapor phase
is exactly equal to its tendency to escape from the vapor phase to the liquid phase.

In the following analysis it is assumed that Tv = Tl = T, Pv = Pl = P and the escaping

tendencies are equal. A special case of the third condition for equilibrium is represented by
Raoults law.

P yi = Pi xi 25.1

Where xi and yi are liquid and vapor mole fractions of component i at temperature T of the
system.

The separation of a binary mixture is represented by a 2D space. The graphical method by

McCabe and Thiele for solution of problems involving binary mixtures is shown in
subsequent section. This method makes use of an equilibrium curve that may be obtained
from the boiling point diagram(BPD).

Construction and interpretation of BPD for binary mixtures

For a binary system having components A and B; the equilibrium relationships are given as
follows : P yA = PA xA , P yB = PA xB, yA + yB = 1, xA + xB = 1. 25.2

Following Gibbs phase rule , the degree of freedom for the above set of equations (2),is 2 and
thus two variables must be fixed in order to get the solution of the equation (2). For
construction of BPD , the total pressure P must be fixed and a solution is obtained for each of
several temperatures lying between the temperatures at which the respective vapor pressures
PA and PB are equal to the total pressure P. The solution of eqation (3.2) for x A in terms of PA
and PB and P is effected as follows P = PAxA + PBxB 25.3

Or, xA = (P-PB)/ (PA - PB) 25.4

From the definition of a mole fraction , equation (4) has a meaningful solution at a given P
for any T lying between the boiling point temperatures TA and TB of pure A and pure B.
Using the first expression of equation (2) yA can also be calculated. By plotting T vs xA and T
vs yA we get the lower and upper curves respectively , fig 1.3 are typical of those obtained
when when component A is more volatile than B. If for the close interval TA TB, PA >
PB , the parallel lines such as CE that join the equilibrium pairs (x,y) , computed at a given
values of T and P by the earlier quations is called Tie lines.

Fig. 25.3: The boiling point Diagram

Determination of bubble point and dew point temperatures of multi-

component mixtures (BPT, DPT)

The state of equilibrium for a two phase system is described by the following equations in
which any number of components c are distributed between the two phases

yi=Ki xi , i= 1 , i= 1 where (1 i c) 25.5

where Ki is ratio of the fugacities of ith component at liquid to vapor phases respectively for
an ideal solution. From this equation it is evident that number of equations is c+2 and
number of variables is 2c+2. Thus, to obtain a solution to these equations , c variables
must be fixed.
When the first expression of Eq(25.5) is summed over for all components , we get

1= i xi 25.6

where Ki is an implicit function of temperature. The solution of the last equation is achieved
by Newtons method . In this method ,the Eq (25.6) is restated as shown

f(T) = i xi 1.

The value of T at which f(T) =0, is the bubble point. In this iterative scheme we need to find
the first derivative of f(T) and further follow the iterative form

Tn+1 = Tn f(Tn)/f(Tn). where Tn is the assumed value for the first stage of the iteration and
Tn+1 is the modified value after the iteration. The process goes on as long as the values of T
are not within a convergence criteria.

When the yis and P are fixed rather than xis and P, the solution temperature of the Eq (25.5)
will give the DBT. Here , just like previous case the governing equation is

F(T)= i/Ki-1.

Kbs method of calculating DBT AND BPT

Robinson and Gilliland pointed out that if the relative values of the K is are independent of
temperature, the Eq (25.5) may be rearranged so that the iterative methods used in previous
case may be avoided. Ratio Ki/Kb is called the relative volatility , i, of component i with
respect to component b; where the K values are calculated at same temperature and
pressure.

For calculation of BPT, Eq(25.5) is rewritten as

yi= (Ki/Kb) Kb xi = I Kb xi 25.7

Taking summation on both sides , we get on rearranging

Kb=1/( i i) 25.8

Since Is are independent of temperature , they maybe computed by Ki & Kb evaluated at

any arbitrary T and at specified pressure. After Kb has been calculated using eq(25.8) , the
BPT can be found from the known relation between Kb and T. The same procedure is
obtained for calculation of DPT.

Separation of Multicomponent Mixtures by use of a Single Equilibrium

Stage

Each of the separation processes considered here are special cases of the general separation
problem in which a multi-component mixture is to be separated into two or more parts
through the use of any number of equilibrium stages.

Flash calculations

The boiling point diagram is useful for visualizing the necessary conditions required for a
flash to occur. Suppose the feed to be flashed has the composition xi , and further suppose
that this liquid must be at the temperature T0 and pressure , P = 1atm, is to be flashed by
raising the temperature to the specified flash temperature TF = T2 at specified flash pressure,
P = 1atm . It is observed that the BPT of the feed TBPT = at P = 1atm is T1. The DPT, TDPT of
the feed at the pressure 1 atm is T3. Then a necessary condition for a flash to occur at the
specified pressure is that ; TBPT < TF < TDPT

In practice, the flash process is generally carried out by reducing the pressure on the feed
stream rather than heating the feed at constant pressure. To determine whether or not the feed
will flash at a given TF & P, the following two conditions must be satisfied:

1. f(TF) > 0 , F(TF) > 0 25.9

2. f(TF) = Fi Xi 1, F(TF) = i/KFi 1 25.10

There are two types of flash calculations : isothermal flash and adiabatic flash.

Isothermal Flash

In isothermal flash the following specifications are made; TF, P, {Xi} and F. We have to find
the unknowns VF, LF, {yFi} and {xFi}. The following equations are needed:

Equilibrium relations : yFi = KFi xFi , , Fi = 1 , Fi = 1 where (1 i c) 25.11

Material balances: Fxi = VF yFi + LF xFi.

Eq(25.10) is seen to represent 2c+2 equations in 2c+2 unknowns. This system of non-linear
equations is readily reduced to one equation in one unknown (say V F) in the following
manner. First observe that the total material balance expression may be obtained by summing
each member of the last expression of Eq(25.9) over all components to give ,

F I = VF Fi + LF Fi or F = VF + LF 25.12

Elimination of yFis from the last expression of Eq(25.9) by use of the first expression ,
followed by rearrangement gives :

xFi = Xi / (LF/F + VFKFV/F) 25.13

Elimination of LF from Eq(25.11) by eq(25.10) yields

xFi = Xi / (1- (1- KFi)) 25.14

where = VF/F.

when each member of Eq(25.12) is summed over all components and the result so obtained is
restated in functional notation , one obtains

P( ) = i/(1- (1- KFi)) 1 25.15

And P( ) = i(1- KFi)/(1- (1- KFi)) 2 25.16

h = -P( )/P( )

Pk+1( ) = Pk( ) + h

The specification of TF implies that the feed either posses precisely the correct amount of
energy for the flash to account at TF at specified P or , that, energy is to be added or
withdrawn at the flash drum.

Enthalpy balance:

FH= VFHF + LFhF 25.17

HF = Fi yFi , hF = Fi XFi 25.18

Adiabatic Flash

The term adiabatic flash is used to describe the problem wherein the following specifications
are made; P,Q = 0 (no heat is added at flash drum) , H, {Xi}, F. In this case there are 2c+3
unknowns [TF, VF,LF,{yFi},{xFi}] and 2c+3 independent equations.

Equilibrium relations, material balances and energy balances are same as in previous case. To
solve this problem using N-R method, it is essential that number of independent functions to
be equal to number of independent variables.

f1 = KF1xF1 yF1

f2 = KF2xF2 yF2

fc = KFcxFc yFc

fc+1 = Fi -1

fc+2 = Fi 1

fc+3 = VFyF1 + LFxF1 F X1

fc+4 = VFyF2 + LFxF2 F X2

f2c+2 = VFyFc + LFxFc F Xc

f2c+3 = VFHF + LFhF FH

The application of the N-R method to this set of equations may be represented by the
following matrix equation : Jx = -f

The jacobian matrix ,J and the column vectors x are defined as follows:

J=

x = [yF1 . yFc xF1 .xFc VF LF TF]T

f = [f1 f2 ..fc fc+1 f2c f2c+1 .. f2c+3]T

where each element of x is equal to the newly predicted values of the variable minus the
assumed value; for example , yF1 = yF1,n+1 - yF1,n. To initiate the calculations , a complete set
of values for the variables must be assumed; say (yF1,n . yFc,n xF1,n .xFc,n VF,n LF,n TF,n).

Alternative process for adiabatic flash:

When the pressure of a liquid stream of known composition , flowrate, and temperature is
reduced adiabatically across a value an adiabatic flash calculation is made to determine the
resulting temperature, compositions and flowrates of equilibrium liquid and vapor streams for
a specified pressure downstream of the value. For an adiabatic flash , the isothermal flash
calculation procedure can be applied in the iterative way. A guess is made of the flash
temperature Tv . Then ,V, x,y and L are determined as for an isothermal flash. The guessed
value of Tv (equal to Tl) is next checked by an energy balance with Q=0 to give

F (Tv) = ( hv + (1- ) hl - hf) /1000 = 0 25.19

Where the division by 1000 is done in order to convert the terms in the order of 1. If the
computed value of F (Tv) is not zero, the entire procedure is repeated for two or more Tv.
The procedure is tedious as it involves inner loop iteration on and outer loop iteration of
Tv. outer loop iteration is successful when equation f( )= 0= (zi)(1 Ki)/(1+ (Ki -
1))

Where = , Ki is a function of {Tv,Pv} and V=

For the closed boiling mixers, the above algorithm dose not work well due to extreme
sensitivity Ts in the inner loop. The above problem can be remove by first assume Tv and
iterated for external loop and inside loop be iterated for .

F(Tv) = (zi)(1 Ki)/(1+ (Ki -1)) 25.20

Then compute x & y from the following equations:-

Xi = zi/ (1+ (Kv-1))

Yi = ziki/ (1+ (Kv-1))=xiki

Eq(1) is solved directly for , from which = (hf - hl)/ (hv - hl) 25.21
If from Eq(25.21) is not equal to assumed then solve Eq(25.20) , the new value of is to
repeat the outer loop starting with Eq(25.20).

Rachford-Rice Procedure for isothermal flash calculations when K-values are independent of
composition follows the algorithm as stated belo

Step 1. Specify variables F, TF, PF, X1,X2, ,Xc, Tv , Pv

Step 2. TL = TV, PL = PV

Step 3. Solve f{ }= (Xi)(1 Ki)/(1+ (Ki -1)), Where = , Ki is a function of

{Tv,Pv} and V=

Step 4. Xi = zi/ (1+ (Kv-1)) , Yi = ziki/ (1+ (Kv-1))=xiki

Step 5. L= F V

Step 6. Q= hvV + hLL hF F is set to 0

Using step 5, and further simplifying , we get hv + (1- )hL hF = 0

We may use softwares like Chemcad design II, PRO/II, ASPEN PLUS to solve these
equations.